Estimation is a division of statistics that determines the values of parameters through measured and observed empirical data. In this paper, we study the estimation of a mean vector restricted to a convex cone with five types of restrictions: positive orthant, simplicial, polyhedral, general convex cones, and nested cones under balanced loss function. We find estimators that improve on the natural estimator in the general case of a spherically symmetric distribution. Also, we introduce the restricted generalized Bayes estimation and two types of the restricted soft threshold wavelet shrinkage estimator by finding Stein’s unbiased risk estimate. Finally, through a simulation study, we investigate the behavior of the restricted soft threshold wavelet shrinkage estimator, and the Parkinson speech dataset is given to test the efficiency of this estimator in denoising. After denoising the real datasets, by computing average mean square error, we find that the proposed estimator dominates other competing estimators.